Unicode in SQLServer

[Nils Jänicke]

While testing the unicode support of TD 5.1 we found a problem that seems to be a router problem with MS SQLServer also on older versions.



create a test database on SQLServer 2005

open sqltalk and connect to the SQLServer

try this script:



create table test (field1 nvarchar(4) primary key, memo ntext);

insert into test (field1) values ('1234');

select field1, memo from test;

select memo, field1 from test;



Look at the output:



FIELD1

======

1234

MEMO

====



1 ROW SELECTED





FIELD1

======

123

MEMO

====



1 ROW SELECTED



You see, that one character of the result of a nvarchar field is dropped when you select a ntext item before, with or without content. This problem occurs only if the field is fully filled.



Nils

[Nils Jänicke]

In addition to that nvarchar columns of SQLServer (connected via ODBC) are not shown correctly in TD 5.1 in the same case. There not only the last character is dropped, sometimes the result is even shorter. Without getting ntext columns it seems to be correct.



Nils

[Jean-Marc Gemperle]

Hi Nils,

Yes can see that currently.... I'll enter a bug asap.

JM

connect ODBSQLS support support;
create table test (field1 nvarchar(4) primary key, memo ntext);
insert into test (field1) values ('1234');
select field1, memo from test;
select memo, field1 from test;
----------------

FIELD1
======
1234
MEMO
====


1 ROW SELECTED






FIELD1
======
1
MEMO
====


1 ROW SELECTED

[Nils Jänicke]

It seems, that this bug is NOT fixed in SP2.

When will I get a version I can use?
Every SP you fix some bugs, but there are already too much bugs, so we can't use it, we spend much time to test and to produce test cases for you but our TD application looks already in battleship gray, so it becomes hard to get new customers and not to loose the old ones.

Nils